Draft Comments Welcome Instrument Choice in Environmental Policy

1 Judging alternative instruments in terms of cost-effectiveness alone, for example, is difficult, since a comprehensive assessment of cost would include not only the negative impacts on the regulated entity but also the costs of monitoring and enforcing the policy in question. It would also include general equilibrium impacts outside of the sector targeted for regulation.

2 See, for example, Keohane, Revesz, and Stavins (1998) for a more in-depth discussion of the (positive) political economy of environmental regulation.

3 Hepburn (2006), Nordhaus (2007), and Tietenberg (2006), among others, provide insightful discussions of some general issues in instrument choice. For discussions with more of an international focus, see, for example, Aldy and Stavins (2007) and Nordhaus (2007).

4 We prefer the term “direct regulatory instruments” to the term “command-and-control” instruments, which has a somewhat negative connotation.

5 As discussed below, certain types of performance standards can also be viewed as technology-promoting policies.

6 This holds even when competitive supply curves are upward sloping and firms cannot pass through all the costs of regulation. In such cases, the emissions price policy causes the producer (net) price to fall. However, the consumer price still exceeds the new producer price by an amount corresponding to the price of emissions.

7 This assumes that eligibility for the subsidy depends on whether the firm remains in business. If the subsidy were fully lump-sum (unconnected with anything other than the reduction in emissions), then it would not affect the firms’ average costs and would not lead to excess entry. See Baumol and Oates (1988) for a discussion.

8 Note that a higher subsidy rate exacerbates the problem of excess entry. As a result, it is possible that, within a range, a higher subsidy rate will lead to higher aggregate emissions. See Baumol and Oates (1988).

9 For example, it may be a lot less costly for firms that are currently upgrading or constructing new plants to incorporate a new abatement technology than for firms that must retrofit older plants that are not readily compatible with the newly mandated technology. According to Deutsch and Moniz (2007), the cost of retrofitting existing coal plants with carbon capture and storage capabilities would greatly exceed that from configuring the technology into the future design and construction of coal plants.

10 A subsidy to encourage widespread adoption of a specific control technology may be even less cost-effective than a mandate to adopt this technology, because the subsidy lowers average production costs below the level under the mandate. This implies higher output than under the technology mandate and an associated loss of cost-effectiveness.

11 For further discussion of the differences between technology mandates and emissions pricing and their implications for cost-effectiveness, see for example Spulber (1985) and Goulder et al. (1999).

12 California has recently introduced a “low carbon fuel standard,” which requires refiners to include a certain minimal percentage of “low carbon” fuel in the motor fuel they sell. At present the only viable option for meeting this standard is blending more ethanol with gasoline.

13 The dominance of the tax interaction effect has been demonstrated in a range of models. For example, Bovenberg and de Mooij (1994) consider a tax on a good related to pollution, where this good is an average substitute for leisure, and environmental damages are separable in the household utility function. Even when pollution tax revenues are recycled in labor tax reductions, the net effect of the tax swap is to reduce labor supply, implying that the tax-interaction effect dominates the revenue-recycling effect. See Bovenberg and Goulder (2002) for a general review of the environmental tax shift literature. The revenue-recycling effect can dominate the tax-interaction effect if the polluting good is a weak enough substitute for leisure, or possibly if reducing an externality has a positive feedback effect on labor supply. It can also dominate if the tax shift serves to reduce other distortions from pre-existing taxes, such as distortions that arise from inefficient relative taxation of capital and labor (Bovenberg and Goulder (1997)) or from tax-deductible consumption expenditure (Parry and Bento (2000)).

14 One can consider an emissions allowance system as equivalent in scale to an emissions tax whose rate equals the market allowance price. If the equivalent allowance system introduces the allowances through an auction, it should generate approximately the same revenue as the equivalently scaled emissions tax, and hence enjoy a comparable revenue-recycling effect. Thus it should involve similar overall costs. Likewise, a system of freely allocated emissions allowances would involve similar costs to an emissions tax system in which the tax revenues are returned to the private sector in a lump-sum fashion, as through fixed rebate checks.

15 Subsidy policies obviously impose a cost in terms of the extra revenue needed from distortionary taxes to finance them (the inverse of the revenue-recycling effect). This can be partly counteracted by a beneficial tax-interaction effect if the policy lowers product prices – as would be the case, for example, with subsidies for clean fuels. But in the case of subsidies for pollution reductions, product prices actually increase as firms are effectively penalized for using inputs. Here the tax-interaction effect flips sign and reinforces the costs of financing the subsidy (Parry (1997).

16 In the context of climate policy, Pizer (2002) compared a pure emissions allowance policy, a global emissions tax, and an emissions allowance system in combination with a safety valve. He found that efficiency gains are highest under the hybrid allowance/safety valve policy, where the ceiling price equals the marginal damage and the emissions cap is fairly tight so that the safety valve is triggered nearly all the time. For further discussion of the safety valve in the context of climate policy, see Burtraw et al. (2002).

17 This abstracts from the various cost complications discussed in Section 3.

18 Uncertainty over the marginal damage curve may modify this result if marginal damages are correlated with marginal costs (Weitzman (1974), Stavins (1996).

19 Pricing rules applying to regulated electric utilities can give rise to an exception to this point. See Burtraw et al. (2006).

20 Note that p_Caef corresponds to a triangle under the marginal abatement cost schedule for the industry with base equal to the reduction in emissions and height equal to the allowance price, while area fedp_S corresponds to a rectangle under the marginal abatement cost schedule with base equal to remaining emissions and height equal to the allowance price.

21 In the first phase of the European Union’s emissions trading program, over 95 percent of the allowances were given away for free. In keeping with the analysis above, this generated windfall profits to many of the regulated firms. Partly in reaction to this result, there has been a distinct shift towards greater emphasis on the auctioning of allowances in the recently established Regional Greenhouse Gas Initiative in the northeast U.S. and in various climate bills recently introduced in the U.S. Congress.

22 Ideally, industry compensation would be progressively phased out over time. This avoids potential difficulties in re-allocating the share of the allowance cap going to different firms, as different firms expand and contract in the future. For example, Stavins (2007) recommends an initial free allocation of half the allowances in a carbon cap-and-trade system, transitioning to full allowance auctions over a 25-year period. This is roughly equivalent in present value terms to granting producers 15-20 percent of allowances in perpetuity.

23 Existing U.S. programs include the sulfur dioxide and nitrogen oxide emissions trading in the Los Angeles region under the Regional Clean Air Incentives Market (RECLAIM) and in the Northeast and Midwest under Title IV of the U.S. Clean Air Act.

24 Household incidence also depends on the distribution of pollution reduction benefits, which depends on the exposure of different income groups to pollution and how those groups value environmental risks. However, for a given reduction in pollution, the distribution of benefits is de-coupled from the choice of emissions control instrument.

25 However, a sticking point is obtaining a satisfactory measure of household income. Economists generally prefer a measure of lifetime or “permanent” income as this better reflects households’ long-run consumption possibilities, though obtaining a reliable measure is difficult when households are constrained in their ability to borrow against expected future earnings. Studies using lifetime income measures generally find that the distribution burden of environmental policies is less regressive than studies that use annual income; that is, the disparity in burden to income ratio between low- and high-income households is reduced.

26 The table’s columns only consider a subset of the dimensions discussed above. Other dimensions – the ease of monitoring and enforcement and the ability to minimize expected policy errors under uncertainty – are also very important, but as discussed above the extent to which a given instrument enjoys these attributes will depend on the specific circumstances involved.

27 For general discussions of policies to simultaneously address pollution and technology market failures see, for example, Newell, Jaffe and Stavins (2003) and Parry (2003).

28 An exception is provided by certain types of performance standards. Such standards can provide significant incentives for invention of new technologies when they are set beyond what is feasible from currently known technologies. Such performance standards are announced in advance of the time they must be met.

29 On the other hand, a “common pool” problem can work toward excessive R&D. This problem stems from the failure of a given firm to account for the fact that its own R&D reduces the likelihood that other firms will obtain innovation rents (Wright 1983). In general, however, this problem appears to be dominated by the appropriability problem.

The problem of suboptimal innovation incentives may be especially severe for environmental technologies, as opposed to commercial technologies for which there is a natural market demand. For example, skepticism over whether the government can make a long-term commitment to emissions pricing, and the desirability of retaining policy discretion to respond to future learning about the severity of environmental threats, undermine the durable and substantial incentives needed for encouraging major investments in emissions-saving technologies.

30 Imposing stiffer emissions prices than warranted by environmental externalities alone—instead of complementing Pigouvian pricing with tailored technology policies—is an inefficient way to promote innovation. Not only does this approach generate excessive short-term abatement but it also fails to differentiate among technologies that may face very different market impediments. For example, alternative automobile fuels and carbon capture and storage technologies might warrant relatively more support than other technologies, to the extent there are network externalities associated with the new pipeline infrastructure required to transport fuels to gas stations, or emissions to underground storage sites.

31 Moreover, to the extent that learning-by-doing provides firm-specific benefits, rather than general benefits to other firms, innovation may occur too rapidly as firms attempt to gain prime-mover advantages at the expense of rival firms.

32 Clearly many if not most economists support the idea that the social rate of discount is lower than the market rate of interest , which implies that, from the point of view of social welfare, consumers tend to discount the future too heavily in their choices of consumer durables or, more broadly, in their saving decisions. This provides a rationale for government support of broad savings incentives rather than incentives focused only on saving in the form of purchases of energy-efficient durable goods.

33 Of course, if it is not possible to deal with each externality through an instrument that focuses on that externality, the choice among feasible instruments – a second-best choice – will depend on the ability of these instruments to deal with the externalities involved. This seems particularly relevant to the choice between gasoline taxes and fuel-economy standards for automobiles. Externalities that vary directly with the amount vehicles are driven – primarily congestion and traffic accidents – are large relative to externalities, such as CO₂ emissions, that vary with the amount of fuel combusted, which in turn depends on both miles driven and fuel economy. Accounting for impacts on mileage-related externalities greatly strengthens the efficiency advantage of higher fuel taxes over fuel economy regulation, even though the main argument for these policies at present is to reduce CO₂ emissions and oil dependence (see, for example, Kleit (2004). This is because higher fuel taxes increase per mile fuel costs, which discourage driving and thereby reduce congestion and accident-related external costs. In contrast, fuel economy regulations reduce per mile fuel costs, and this tends to promote more driving – a “rebound” effect.

Directory: economics
economics -> Macroeconomics Machine-graded Assessment Items Module: Macroeconomic Measures of Performance
economics -> Critical habitat designation for the california gnatcatcher
economics -> Chapter 7: The European Union (latest revision 2009) Development of the European Union
economics -> Social Network Construction and Competitiveness in Global Value Chain Network Analysis Perspective on Chinese Aerospace Industry
economics -> San José State University Department of Economics Econ 235—Seminar in Monetary Theory and Policy
economics -> General Introduction to Market Structures The worlds cheapest tablet computer is the Aakash tablet. Why is it so cheap?
economics -> Instructions:  This is a 120-minute examination
economics -> Contents Introduction
economics -> Nato and its future Have combat experience, will travel
economics -> Economics 102 Assignment #1 (15 Points) Name